8ft) Mr. J. C. McConnel. [M.ir. 12. 



bent faces throughout. If the black centre was near tbe middle of 

 one half, the sharp bend was crowded with narrow coloured bands 

 which moved slowly along as the bar was tilted, till as each band 

 reached the straight piece beyond the bend it moved rapidly and 

 broadened out. 



The movement of the bands across the bend, though slow, was quite 

 regular, so the direction of the optic axis changed quickly but not 

 per salt tint. 



Exp. 4. A similar arrangement. The bar was all one crystal except 

 the parts actually on the supports. The optic axis was transverse, 

 but horizontal. Depth, 9'5 mm. ; breadth, 10 mm. ; supports, 75 mm. 

 apart. The weight of 1'29 kilos, was applied over 42 hours from 

 4.15 P.M. on December 15 to 10.35 A.M. on December 17. The 

 minimum temperature was 7'8 C., the maximum l'l C., the mean 

 about -3'3 C. 



Decided evaporation had taken place ; the edges of the bar wore 

 rounded :and the string which had stuck to the bar was raised on 

 ridges. The greater part of the bar was 8^ mm. deep, 9 mm. broad. 

 In comparing the traces taken before and after the experiment 

 I could find no bending. It certainly did not amount to half a milli- 

 metre. The traces were taken by laying the bar on a sheet of 

 paper and following the upper and lower edges with a pencil. 



Exp. 5. The same bar, turned so as to put the optic axis vertical, 

 bent rapidly. 



Depth 8f mm., breadth 8j mm. Distance between supports 73mm. 

 The weight of 0'62 kilo, was applied from 11.10 A.M. to 9.5 P.M. on 

 December 17. The minimum temperature was 4'4 C., the maxi- 

 mum 1'7 C., mean about 3'0. The depression of the middle J 

 measured on the trace was about 4'4 mm., which had taken place \ 

 in 10 hours. Assuming that in Experiment 4 the depression was less 

 then 0'5 mm., the bending of the bar in the new position must ha\ 

 been at least thirty-seven times as fast. It is true the depth and 

 breadth were slightly less, but the weight was less than half as great. 

 The results of 'Exp. 3 as to bubbles and optic axis were confirmed. 



Exp. 6. A bar with the axis longitudinal. 



I obtained a large lump of thick ice from the Davos lake, and 

 from this cut a <bar which appeared to be all one crystal, with the 

 axis longitudinal. I need not enter on the details of the experiment, 

 especially as the temperature rose above freezing-point. But at the 

 end the bar had the shape shown in the diagram, Figs. 3, 4. 



The dotted line indicates a division between the crystals. The 

 double-headed arrows show the direction of the optic axis in different 

 parts, or at least the projection of that direction on the plane of the 

 paper. This was determined by making the field of the polariscope 

 dark as possible, putting the part of the bar in question in the middle 



